Hackerschool AY 20/21 S 1: Markov Chain Monte Carlo

Markov chain Monte Carlo (MCMC) methods were invented around the time of World War II by several mathematicians and physicists who worked on the Manhattan Project. What started out as a way to solve high-dimensional integrals from chemical physics is now a general, powerful approach to generate samples from probability distributions. It is widely applied in statistics, computations for the natural sciences, and more recently machine learning and data science. Regarded as one of the top 10 most important algorithms in the 20th century, MCMC is undoubtedly an essential tool in any scientist’s technical repertoire.

In this workshop, we will learn about the keys ideas of MCMC, bits of mathematical theory, and its applications and implementations, including variants such as Random Walk Metropolis and Hamiltonian Monte Carlo.